More Time for Requests to Finish During Workers Runtime Updates: Up to 20 Minutes

Hi everyone, it's Shii-chan!

In-flight requests now get much more time to finish during Workers runtime updates!

What was announced?

The Cloudflare Changelog says pending requests and connections to Workers now have a grace period of up to 20 minutes during runtime updates.

The grace period used to be up to 30 seconds, so long-running requests could be cut off depending on update timing.

These kinds of work have more time to continue before an update interrupts them:

Here's how an update behaves:

New incoming HTTP requests go to the updated runtime once it is ready
Processes running the previous version stop accepting these requests and drain existing work
The grace period applies to each process running the previous version
When it ends, unfinished requests and open connections on that process are terminated

A longer grace period doesn't mean interruptions are gone:

Design your app to reconnect WebSockets, safely retry interrupted requests, and recover saved state after a restart
WebSockets can disconnect before the grace period ends, including connections using the WebSocket Hibernation API

See the Workers runtime updates docs for more.

The grace period grew from up to 30 seconds to up to 20 minutes
Long HTTP requests, AI streams, and WebSockets have more room to finish
Interruptions can still happen, so keep reconnect, retry, and state recovery in your design

If you run long-lived connections or AI streaming on Workers, this one's for you!